From: Boris Pismenny <borispismenny@gmail.com> To: Or Gerlitz <gerlitz.or@gmail.com>, David Ahern <dsahern@gmail.com> Cc: Boris Pismenny <borisp@mellanox.com>, Jakub Kicinski <kuba@kernel.org>, David Miller <davem@davemloft.net>, Saeed Mahameed <saeedm@nvidia.com>, Christoph Hellwig <hch@lst.de>, Sagi Grimberg <sagi@grimberg.me>, axboe@fb.com, Keith Busch <kbusch@kernel.org>, Alexander Viro <viro@zeniv.linux.org.uk>, Eric Dumazet <edumazet@google.com>, smalin@marvell.com, Yoray Zack <yorayz@mellanox.com>, yorayz@nvidia.com, boris.pismenny@gmail.com, Ben Ben-Ishay <benishay@mellanox.com>, benishay@nvidia.com, linux-nvme@lists.infradead.org, Linux Netdev List <netdev@vger.kernel.org>, Or Gerlitz <ogerlitz@mellanox.com>, Or Gerlitz <ogerlitz@nvidia.com> Subject: Re: [PATCH v4 net-next 06/21] nvme-tcp: Add DDP offload control path Date: Sun, 21 Feb 2021 13:29:39 +0200 [thread overview] Message-ID: <3f447ef7-3ded-8c1c-fdfa-ec137e11a19a@gmail.com> (raw) In-Reply-To: <CAJ3xEMjMqK81uNv21poD+DoZCRxYak0DYFZrjDbmWaSxw4R5ig@mail.gmail.com> On 17/02/2021 15:55, Or Gerlitz wrote: > On Sun, Feb 14, 2021 at 8:20 PM David Ahern <dsahern@gmail.com> wrote: >> On 2/11/21 2:10 PM, Boris Pismenny wrote: >>> @@ -223,6 +229,164 @@ static inline size_t nvme_tcp_pdu_last_send(struct nvme_tcp_request *req, >>> return nvme_tcp_pdu_data_left(req) <= len; >>> } >>> >>> +#ifdef CONFIG_TCP_DDP >>> + >>> +static bool nvme_tcp_resync_request(struct sock *sk, u32 seq, u32 flags); >>> +static const struct tcp_ddp_ulp_ops nvme_tcp_ddp_ulp_ops = { >>> + .resync_request = nvme_tcp_resync_request, >>> +}; >>> + >>> +static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue) >>> +{ >>> + struct net_device *netdev = queue->ctrl->offloading_netdev; >>> + struct nvme_tcp_ddp_config config = {}; >>> + int ret; >>> + >>> + if (!(netdev->features & NETIF_F_HW_TCP_DDP)) >> >> If nvme_tcp_offload_limits does not find a dst_entry on the socket then >> offloading_netdev may not NULL at this point. > > correct :( will look on that > That's only partially true. If nvme_tcp_offload_limits finds a dst_entry, but then the netdevice goes down, then the check here will catch it. This is needed because nvme_tcp_offload_limits doesn't hold a reference! We opted not to grab a reference on nvme_tcp_offload_limits because it doesn't create a context. >>> + queue->ctrl->offloading_netdev = NULL; >>> + return -ENODEV; >>> + } >>> + >>> + if (netdev->features & NETIF_F_HW_TCP_DDP && >>> + netdev->tcp_ddp_ops && >>> + netdev->tcp_ddp_ops->tcp_ddp_limits) >>> + ret = netdev->tcp_ddp_ops->tcp_ddp_limits(netdev, &limits); >>> + else >>> + ret = -EOPNOTSUPP; >>> + >>> + if (!ret) { >>> + queue->ctrl->offloading_netdev = netdev; >> >> you save a reference to the netdev here, but then release the refcnt >> below. That device could be deleted between this point in time and the >> initialization of all queues. > That's true, and this is why we repeat the checks there. We avoid holding the reference here because there is no obvious complementary release point for nvme_tcp_offload_limits and there is no hardware context created here, so there is no real need to hold it at this stage.
WARNING: multiple messages have this Message-ID (diff)
From: Boris Pismenny <borispismenny@gmail.com> To: Or Gerlitz <gerlitz.or@gmail.com>, David Ahern <dsahern@gmail.com> Cc: smalin@marvell.com, linux-nvme@lists.infradead.org, Sagi Grimberg <sagi@grimberg.me>, yorayz@nvidia.com, boris.pismenny@gmail.com, Ben Ben-Ishay <benishay@mellanox.com>, Yoray Zack <yorayz@mellanox.com>, benishay@nvidia.com, David Miller <davem@davemloft.net>, axboe@fb.com, Eric Dumazet <edumazet@google.com>, Alexander Viro <viro@zeniv.linux.org.uk>, Linux Netdev List <netdev@vger.kernel.org>, Keith Busch <kbusch@kernel.org>, Jakub Kicinski <kuba@kernel.org>, Or Gerlitz <ogerlitz@mellanox.com>, Or Gerlitz <ogerlitz@nvidia.com>, Saeed Mahameed <saeedm@nvidia.com>, Christoph Hellwig <hch@lst.de>, Boris Pismenny <borisp@mellanox.com> Subject: Re: [PATCH v4 net-next 06/21] nvme-tcp: Add DDP offload control path Date: Sun, 21 Feb 2021 13:29:39 +0200 [thread overview] Message-ID: <3f447ef7-3ded-8c1c-fdfa-ec137e11a19a@gmail.com> (raw) In-Reply-To: <CAJ3xEMjMqK81uNv21poD+DoZCRxYak0DYFZrjDbmWaSxw4R5ig@mail.gmail.com> On 17/02/2021 15:55, Or Gerlitz wrote: > On Sun, Feb 14, 2021 at 8:20 PM David Ahern <dsahern@gmail.com> wrote: >> On 2/11/21 2:10 PM, Boris Pismenny wrote: >>> @@ -223,6 +229,164 @@ static inline size_t nvme_tcp_pdu_last_send(struct nvme_tcp_request *req, >>> return nvme_tcp_pdu_data_left(req) <= len; >>> } >>> >>> +#ifdef CONFIG_TCP_DDP >>> + >>> +static bool nvme_tcp_resync_request(struct sock *sk, u32 seq, u32 flags); >>> +static const struct tcp_ddp_ulp_ops nvme_tcp_ddp_ulp_ops = { >>> + .resync_request = nvme_tcp_resync_request, >>> +}; >>> + >>> +static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue) >>> +{ >>> + struct net_device *netdev = queue->ctrl->offloading_netdev; >>> + struct nvme_tcp_ddp_config config = {}; >>> + int ret; >>> + >>> + if (!(netdev->features & NETIF_F_HW_TCP_DDP)) >> >> If nvme_tcp_offload_limits does not find a dst_entry on the socket then >> offloading_netdev may not NULL at this point. > > correct :( will look on that > That's only partially true. If nvme_tcp_offload_limits finds a dst_entry, but then the netdevice goes down, then the check here will catch it. This is needed because nvme_tcp_offload_limits doesn't hold a reference! We opted not to grab a reference on nvme_tcp_offload_limits because it doesn't create a context. >>> + queue->ctrl->offloading_netdev = NULL; >>> + return -ENODEV; >>> + } >>> + >>> + if (netdev->features & NETIF_F_HW_TCP_DDP && >>> + netdev->tcp_ddp_ops && >>> + netdev->tcp_ddp_ops->tcp_ddp_limits) >>> + ret = netdev->tcp_ddp_ops->tcp_ddp_limits(netdev, &limits); >>> + else >>> + ret = -EOPNOTSUPP; >>> + >>> + if (!ret) { >>> + queue->ctrl->offloading_netdev = netdev; >> >> you save a reference to the netdev here, but then release the refcnt >> below. That device could be deleted between this point in time and the >> initialization of all queues. > That's true, and this is why we repeat the checks there. We avoid holding the reference here because there is no obvious complementary release point for nvme_tcp_offload_limits and there is no hardware context created here, so there is no real need to hold it at this stage. _______________________________________________ Linux-nvme mailing list Linux-nvme@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-nvme
next prev parent reply other threads:[~2021-02-21 11:30 UTC|newest] Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top 2021-02-11 21:10 [PATCH v4 net-next 00/21] nvme-tcp receive offloads Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 01/21] net: Introduce direct data placement tcp offload Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 02/21] net: Introduce crc offload for tcp ddp ulp Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 03/21] iov_iter: DDP copy to iter/pages Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 04/21] net: skb copy(+hash) iterators for DDP offloads Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 05/21] net/tls: expose get_netdev_for_sock Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 06/21] nvme-tcp: Add DDP offload control path Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-14 18:16 ` David Ahern 2021-02-14 18:16 ` David Ahern 2021-02-15 7:57 ` Or Gerlitz 2021-02-15 7:57 ` Or Gerlitz 2021-02-17 13:55 ` Or Gerlitz 2021-02-17 13:55 ` Or Gerlitz 2021-02-21 11:29 ` Boris Pismenny [this message] 2021-02-21 11:29 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 07/21] nvme-tcp: Add DDP data-path Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-14 18:27 ` David Ahern 2021-02-14 18:27 ` David Ahern 2021-02-17 14:01 ` Or Gerlitz 2021-02-17 14:01 ` Or Gerlitz 2021-02-17 17:00 ` David Ahern 2021-02-17 17:00 ` David Ahern 2021-02-21 11:44 ` Boris Pismenny 2021-02-21 11:44 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 08/21] nvme-tcp: RX CRC offload Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 09/21] nvme-tcp: Deal with netdevice DOWN events Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 10/21] net/mlx5: Header file changes for nvme-tcp offload Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 11/21] net/mlx5: Add 128B CQE for NVMEoTCP offload Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 12/21] net/mlx5e: TCP flow steering for nvme-tcp Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 13/21] net/mlx5e: NVMEoTCP offload initialization Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 14/21] net/mlx5e: KLM UMR helper macros Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 15/21] net/mlx5e: NVMEoTCP use KLM UMRs Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 16/21] net/mlx5e: NVMEoTCP queue init/teardown Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 17/21] net/mlx5e: NVMEoTCP async ddp invalidation Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 18/21] net/mlx5e: NVMEoTCP ddp setup and resync Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 19/21] net/mlx5e: NVMEoTCP, data-path for DDP+CRC offload Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:10 ` [PATCH v4 net-next 20/21] net/mlx5e: NVMEoTCP statistics Boris Pismenny 2021-02-11 21:10 ` Boris Pismenny 2021-02-11 21:32 ` [PATCH v4 net-next 00/21] nvme-tcp receive offloads Randy Dunlap 2021-02-11 21:32 ` Randy Dunlap 2021-02-12 5:22 ` Boris Pismenny 2021-02-12 5:22 ` Boris Pismenny 2021-02-21 8:52 ` Or Gerlitz 2021-02-21 8:52 ` Or Gerlitz
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=3f447ef7-3ded-8c1c-fdfa-ec137e11a19a@gmail.com \ --to=borispismenny@gmail.com \ --cc=axboe@fb.com \ --cc=benishay@mellanox.com \ --cc=benishay@nvidia.com \ --cc=boris.pismenny@gmail.com \ --cc=borisp@mellanox.com \ --cc=davem@davemloft.net \ --cc=dsahern@gmail.com \ --cc=edumazet@google.com \ --cc=gerlitz.or@gmail.com \ --cc=hch@lst.de \ --cc=kbusch@kernel.org \ --cc=kuba@kernel.org \ --cc=linux-nvme@lists.infradead.org \ --cc=netdev@vger.kernel.org \ --cc=ogerlitz@mellanox.com \ --cc=ogerlitz@nvidia.com \ --cc=saeedm@nvidia.com \ --cc=sagi@grimberg.me \ --cc=smalin@marvell.com \ --cc=viro@zeniv.linux.org.uk \ --cc=yorayz@mellanox.com \ --cc=yorayz@nvidia.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.